If you check
on the latest tech trends, chatbots have been a lot in the news as the next big
thing. There has been a lot of hype about how successful they would become. Not
to mention the speculation of how they would take the world by storm. But this
yet to happen! In today’s post, we will look at why chatbots did not live to
become as popular as we expected.
But first things first.
is a chatbot?
A bot is a computer program. It helps automate routine tasks by chatting with a user via a conversational interface. Chatbots are powered by artificial intelligence. They can therefore understand complex requests and offer personalized responses and enhance interactions over time.
of rules that govern most chatbots are set by humans through a Bot building
platform. Developers play a crucial role in determining how each of the
conversations is scripted, and the kind of experience users should expect when
interacting with bots.
do they fail?
sometimes fail to deliver user experiences that are as efficient, seamless and
delightful as wished-for. Here are some of the reasons why chatbots fail or may
not be as popular as you may think:
Artificial intelligence is not that accessible
chatbots are not actually intelligent. Since they are created based on a
decision tree logic, they base their responses on specific keywords identified
in the user’s input. This simply means that the intelligence of these types of bots
depends on the capacity, thoughtfulness, and patience of the programmer or
designer who developed it to predict all potential user use cases and inputs.
a developer takes time to think of every possible scenario, life might still
fail to fit into those boxes. To add to that, bots with natural language
learning and linguistic capabilities are still rare.
Use cases are not that strong
new technology is put out in the world, developers and designers get really
excited about it. Chatbots are no exception. What we saw when bots were first
introduced was a gold rush of companies doing their best to be the first in
their category to successfully deploy a bot. The result is an excess of bots
solving irrelevant problems or offering poor experiences.
We have to
learn a lot from our failures before we can deploy smart and relevant bots. Before developing a bot, it would be useful to
answer the following questions:
product really need a bot?
Do I have
the patience to build a bot that will do exactly what I want it to do?
platforms that can support its functionality?
developers tend to bypass such questions. No wonder most these bots never
Lack of transparency
chatbots are transparent. Right from the beginning, they let the user know that
they are chatting with a robot, not a human being. Knowing that they are
chatting with a machine makes the users more forgiving about some of the mistakes
that the bots may make.
all want to use bots that feel as human as possible, you don’t want to deceive
your users. Pretending like your bot is human will set unrealistic expectations
in the user’s mind and eventually lead to loss of trust when these expectations
are not met. Try chatting with our bot and
let us know if you found it efficient.
Bots don’t understand context
humans good at conversations? They read between the lines, understand sarcasm,
and constantly leverage contextual information when they give you a response.
not have this capacity. Unless in cases where they are powered by natural
language processing technology, they can only hold contextual information for a
few chat bubbles. They end up losing track of what the user said before they
posed the question.
They don’t communicate with existing business
chatbots, designers are often tempted to rebuild functionality from scratch. Bots
should be created as part of a larger ecosystem. Isolating it from the other
business systems can be harmful to customers as well as your business.
example, let’s take the case of a bot built for booking appointments in an
office. The chatbot should have a way of communicating with the offices existing
appointment management system. Otherwise, the office manager will have extra
work trying to handle requests coming through the new channel. It will also
mean lack of consistency for the user.
· Chatbots try to handle too many things at once
It looks like designers are so excited about
the numerous tasks a bot can help execute that they forget to narrow down its
area of focus.
instance, reports indicate that 70% of the bots on Facebook messenger fail at
fulfilling their simple user requests. This is usually as a result of designers
failing to narrow their bot down to one area of focus.
should keep in mind that it would be better to have bots doing one thing well
that to have bots doing multiple things poorly.
· Chatbots lack proper human escalation protocols
to know that they can still rely on humans whenever technology fails. However,
most chatbots do not have an escalation workflow to allow a human to take over
the conversation when the bot is unable to help. They therefore end up leaving users hanging
and sometimes, more frustrated than they were when they started the
People still prefer to talk to other people
As we highlighted
earlier, conversations are made up of way more than text. Bots are not able to
keep up with such things as sarcasm and context. Even bots that are fired by
NLP can only go as far as producing processed content, but they still cannot
compare to humans. No matter how much
wit and human-like mannerisms are incorporated into a bot, it will still be no
match for a human.
and done. Innovators were not entirely wrong. We can continue to use bots to
help us with repetitive, low level and automated queries and tasks. They can
still function as cogs in much larger and complex systems.
chatbots are nowhere close to the expectations created by the 21st
century hype. Computers are simply not good at understanding human emotion.
They are still unable to understand how we feel or even what we are asking them.
thoughts with us here at curator.io.